The long-term goal of this project is to increase the cure rate of primary liver cancer by adapting treatment to the individual patient response. This goal will be carried out through two specific aims. Aim 1 is to complete a phase II clinical trial for patients with HCC in which the dose is based on the individual patient's liver function. We will measure global and regional changes in liver function in patients undergoing radiation therapy for HCC who are ineligible for or have progressed after standard therapy. We will administer 60% of the treatment, wait the month that we have shown is required to express early liver injury, reassess tumor response, liver function, and liver injury to adapt the dose for the rest of treatment, based on a more accurate estimate of that individual patient's risk of toxicity. We will also assess regional response in the tumor. We hypothesize that adapting radiation therapy based on the tumor and uninvolved liver responses assessed during treatment will permit us to achieve a >80% 1 year tumor control rate while causing no greater toxicity than standard therapy. Aim 2 is to conduct a phase II randomized trial to compare standard therapy (fixed course treatment without adaptation) to the model proposed in Aim 1 that adapts therapy based on the individual patient's response. Aim 1 will provide the data to Project 4 to help develop a model for treatment that is predicted to maximize tumor response by intensifying treatment to refractory subvolumes of tumor (defined by Project 3), and minimize uninvolved liver injuty by delivering dose through non-functioning regions of the liver, and by knowing the radiation sensitivity of the individual patient. This will be accomplished by replanning treatment after 60% has been delivered based on tumor response and uninvolved liver injury. This arm will be compared in a randomized phase 11 trial to a standard arm using an approach that bases the dose on the volume of uninvolved liver spared (which we developed in an earlier version of this PO-1), but adapts neither tumor nor uninvolved liver dose and dose distribution. We hypothesize that treating patients with adaptive radiation therapy based on the tumor and uninvolved liver responses assessed during treatment will produce superior tumor control at no greater toxicity than the standard arm that gives a fixed dose based on the volume of uninvolved liver irradiated (but does not adapt therapy). The impact of this work is that it will develop a new paradigm for radiation oncology in which treatment is optimized by adapting therapy to the individual patient response, rather than based on the average patient.